Security context
What an agent needs to avoid regressing past fixes and find the next vuln in this repo.
confident-ai/deepeval
main @ 8bb5b02
2
Fixes
0
CVEs
MEDIUM
Peak severity
Highlights
Information Disclosure: 2 prior fixes. Scrutinize any change in this area.
deepeval/models/llms/anthropic_model.py: most-fixed (1 issue). Treat as high-risk during review.
0 high-severity fixes in this history; regressions here are high-impact.
Recurring patterns
The bug types that recur here, drawn from past fixes, not open vulnerabilities.
Information Disclosure: The Anthropic API key was initialized as a plaintext attribute, exposing it to potential exposure in logs, exception stack traces, and standard model serialization. It was remediated by wrapping the key in a Pydantic SecretStr.
Information Disclosure: The Moonshot/Kimi API key was vulnerable to serialization leaks and accidental exposure through standard output streams. Ensuring the key is wrapped within a SecretStr container prevents direct string interpolation in debug outputs.